Jörg Friedrichs
Associate Professor of Politics, University of Oxford
AI governance should be examined through the lens of infrastructure. From the steam engine to electricity and the internet, every disruptive technological revolution in history has undergone an evolution from invention through adoption and diffusion to infrastructure development and institutionalized regulation. The same applies to AI: large language models themselves constitute a form of infrastructure, while cloud services, data centers, fiber-optic networks, and power supply form the material foundation supporting AI operations. The rapid expansion of data centers is generating enormous energy demands, placing significant pressure on the power systems of multiple countries. Although technology companies dominate technical development, they remain highly dependent on public infrastructure such as energy and network access provided or controlled by states, compelling governments to assume the task of infrastructure governance. Digital platforms represent another infrastructure needed by the open-source coding community. Therefore, AI governance should not focus solely on technical design or ethical dimensions, but should also place greater emphasis on institutional development and standard-setting at the infrastructure level.
